Jürgen R. Reuter, DESY. J.R.Reuter Status Report on WHIZARD ALCW 2015, KEK,

Transcription

1 Status Report on the Event Generator WHIZARD Jürgen R. Reuter, DESY

2 The WHIZARD Event Generator Universal event generator for lepton and hadron colliders Modular package: - Phase space parameterization (resonances, collinear emission, Coulomb etc.) - O Mega optimized matrix element generator (tree level, NLO external) - VAMP: adaptive multi-channel Monte Carlo integrator - CIRCE1/2: generator/simulation tool for lepton collider beam spectra - Modules for beam structure, parton shower, matching/merging, event formats, analysis, cascade decays, polarized initial/final states, [NLO subtractions] etc. - Interfaces to external packages for Feynman rules, hadronization, tau decays, event formats, analysis, jet clustering etc. - SINDARIN: free-format steering language for all inputs (!)

3 The WHIZARD Event Generator Universal event generator for lepton and hadron colliders Modular package: - Phase space parameterization (resonances, collinear emission, Coulomb etc.) - O Mega optimized matrix element generator (tree level, NLO external) - VAMP: adaptive multi-channel Monte Carlo integrator - CIRCE1/2: generator/simulation tool for lepton collider beam spectra - Modules for beam structure, parton shower, matching/merging, event formats, analysis, cascade decays, polarized initial/final states, [NLO subtractions] etc. - Interfaces to external packages for Feynman rules, hadronization, tau decays, event formats, analysis, jet clustering etc. - SINDARIN: free-format steering language for all inputs (!) v1.0 Project start ca (parts early 90 s): TESLA studies W, Z, Higgs (+ resp. decays) v /2002: optimized matrix elements (O Mega) v /2003: first ever multi-leg implementation of the MSSM v /2006: QCD color flow formalism v1.95/97 02/2010: NMSSM, UED, parton shower (alpha), development stop v1 v /2010: OO overhaul (38 months), modern v2 version, faster matrix elements v /2012: FSR/ISR shower, SINDARIN, unit tests etc., cascade processes v /2014: 2nd OO overhaul (18 months) v /2015: production version, LCIO, NLO alpha, POWHEG alpha, top threshold

4 WHIZARD: Some (technical) facts WHIZARD v2.2.5 ( ) WHIZARD Team: Wolfgang Kilian, Thorsten Ohl, JRR Bijan Chokoufé/Marco Sekulla/Christian Weiss + 2 Master + 2 PhD (soon) (some losses: C. Speckner [software engineering], F. Bach [ESA Space Defense], S. Schmidt [Philosophy]) Publication: EPJ C71 (2011) 1742 (and others for O Mega, Interfaces, color flow formalism)

5 WHIZARD: Some (technical) facts WHIZARD v2.2.5 ( ) WHIZARD Team: Wolfgang Kilian,Thorsten Ohl, JRR Bijan Chokoufé/Marco Sekulla/Christian Weiss + 2 Master + 2 PhD (soon) (some losses: C. Speckner [software engineering], F. Bach [ESA Space Defense], S. Schmidt [Philosophy]) Publication: EPJ C71 (2011) 1742 (and others for O Mega, Interfaces, color flow formalism) 2nd WHIZARD Workshop Würzburg, 03/2015 J.R.Reuter Status Report on WHIZARD ALCW 2015, KEK,

6 WHIZARD: Some (technical) facts WHIZARD v2.2.5 ( ) WHIZARD Team: Wolfgang Kilian,Thorsten Ohl, JRR Bijan Chokoufé/Marco Sekulla/Christian Weiss + 2 Master + 2 PhD (soon) (some losses: C. Speckner [software engineering], F. Bach [ESA Space Defense], S. Schmidt [Philosophy]) Publication: EPJ C71 (2011) 1742 (and others for O Mega, Interfaces, color flow formalism) 2nd WHIZARD Workshop Würzburg, 03/2015 J.R.Reuter Status Report on WHIZARD support junior developers ALCW 2015, KEK,

7 WHIZARD: Some (technical) facts Programming Languanges: Fortran2003/2008 (gfortran 4.7.4), OCaml ( ) Standard conformance to autotools: libtool/autoconf/automake Standard installation: configure <FLAGS>, make, [make check], make install Modern OO programming: abstract modules, polymorphism, inheritance etc. etc. Version control system Hepforge), internal ticket system Large self test suite, unit tests [module tests], regression testing Continous integration system Siegen) NEW: ticketing system DESY) for user support issues (questions, bug reports)

8 WHIZARD: Some (technical) facts Programming Languanges: Fortran2003/2008 (gfortran 4.7.4), OCaml ( ) Standard conformance to autotools: libtool/autoconf/automake Standard installation: configure <FLAGS>, make, [make check], make install Modern OO programming: abstract modules, polymorphism, inheritance etc. etc. Version control system Hepforge), internal ticket system Large self test suite, unit tests [module tests], regression testing Continous integration system Siegen) NEW: ticketing system DESY) for user support issues (questions, bug reports) WHIZARD Hepforge

9 WHIZARD: Some (technical) facts Programming Languanges: Fortran2003/2008 (gfortran 4.7.4), OCaml ( ) Standard conformance to autotools: libtool/autoconf/automake Standard installation: configure <FLAGS>, make, [make check], make install Modern OO programming: abstract modules, polymorphism, inheritance etc. etc. Version control system Hepforge), internal ticket system Large self test suite, unit tests [module tests], regression testing Continous integration system Siegen) NEW: ticketing system DESY) for user support issues (questions, bug reports) WHIZARD Hepforge Talk concentrates on NEW features and current developments/ (near) future plans

10 General structure of SINDARIN input model = SM alias ll = e- : e+ : mu+ : mu- alias nu = n1:n1:n2:n2:n3:n3 alias jet = u:u:d:d:s:s:g process tth = e1, E1 => t, tbar, h process tthfull = e1, E1 => ll, nu, ll, nu, b, bbar, jet, jet process inclusive = e1, E1 => (Z, h) + (Z, Z) + (Wp, Wm) process t_dec = t => E1, nubar, b sqrts = 500 GeV beams = e1, E1 => circe1 => ISR cuts = all M > 10 GeV [jet, jet] integrate (tthfull) { iterations = 15:500000, 5: } n_events = unstable t (t_dec) sample_format = lhef, stdhep, hepmc sample = mydata

11 General structure of SINDARIN input model = SM LCWS 14, Belgrade, Simulation summary talk: WHIZARD Task to implement LCIO format alias ll = e- : e+ : mu+ : mu- alias nu = n1:n1:n2:n2:n3:n3 alias jet = u:u:d:d:s:s:g process tth = e1, E1 => t, tbar, h process tthfull = e1, E1 => ll, nu, ll, nu, b, bbar, jet, jet process inclusive = e1, E1 => (Z, h) + (Z, Z) + (Wp, Wm) process t_dec = t => E1, nubar, b sqrts = 500 GeV beams = e1, E1 => circe1 => ISR cuts = all M > 10 GeV [jet, jet] integrate (tthfull) { iterations = 15:500000, 5: } n_events = unstable t (t_dec) sample_format = lhef, stdhep, hepmc sample = mydata

12 General structure of SINDARIN input model = SM alias ll = e- : e+ : mu+ : mu- alias nu = n1:n1:n2:n2:n3:n3 alias jet = u:u:d:d:s:s:g process tth = e1, E1 => t, tbar, h process tthfull = e1, E1 => ll, nu, ll, nu, b, bbar, jet, jet process inclusive = e1, E1 => (Z, h) + (Z, Z) + (Wp, Wm) process t_dec = t => E1, nubar, b LCWS 14, Belgrade, Simulation summary talk: WHIZARD Task to implement LCIO format WHIZARD v2.2.4, 02/2015: sample_format = lcio simulate (<process>) sqrts = 500 GeV beams = e1, E1 => circe1 => ISR cuts = all M > 10 GeV [jet, jet] integrate (tthfull) { iterations = 15:500000, 5: } n_events = unstable t (t_dec) sample_format = lhef, stdhep, hepmc sample = mydata

13 General structure of SINDARIN input model = SM alias ll = e- : e+ : mu+ : mu- alias nu = n1:n1:n2:n2:n3:n3 alias jet = u:u:d:d:s:s:g process tth = e1, E1 => t, tbar, h process tthfull = e1, E1 => ll, nu, ll, nu, b, bbar, jet, jet process inclusive = e1, E1 => (Z, h) + (Z, Z) + (Wp, Wm) process t_dec = t => E1, nubar, b LCWS 14, Belgrade, Simulation summary talk: WHIZARD Task to implement LCIO format WHIZARD v2.2.4, 02/2015: sample_format = lcio simulate (<process>) sqrts = 500 GeV beams = e1, E1 => circe1 => ISR cuts = all M > 10 GeV [jet, jet] integrate (tthfull) { iterations = 15:500000, 5: } n_events = unstable t (t_dec) sample_format = lhef, stdhep, hepmc sample = mydata

14 Lepton Collider Beam Simulation Tesla, s =500GeV Another demand: adapt GuineaPig beam spectra for WHIZARD v2 For WHIZARD v1.95 simulations done by Lumilinker [T. Barklow] TESLA/SLC spectra were rather simple Fits with 6 or 7 parameters possible [CIRCE1] Beams not factorizable: No simple power law: D B1 B 2 (x 1,x 2 ) 6= D B1 (x 1 ) D B2 (x 2 ) D B1 B 2 (x 1,x 2 ) 6= x 1 1 (1 x 1) 1 x 2 2 (1 x 2) 2 x e ± = x e ±

15 Lepton Collider Beam Simulation Tesla, s =500GeV Another demand: adapt GuineaPig beam spectra for WHIZARD v2 For WHIZARD v1.95 simulations done by Lumilinker [T. Barklow] TESLA/SLC spectra were rather simple Fits with 6 or 7 parameters possible [CIRCE1] Beams not factorizable: No simple power law: D B1 B 2 (x 1,x 2 ) 6= D B1 (x 1 ) D B2 (x 2 ) D B1 B 2 (x 1,x 2 ) 6= x 1 1 (1 x 1) 1 x 2 2 (1 x 2) 2 x e ± = x e ± Dalena/Esbjerg/Schulte [LCWS 2011] CLIC much more complicated (wakefields)

16 Lepton Collider Beam Simulation Tesla, s =500GeV Another demand: adapt GuineaPig beam spectra for WHIZARD v2 For WHIZARD v1.95 simulations done by Lumilinker [T. Barklow] TESLA/SLC spectra were rather simple Fits with 6 or 7 parameters possible [CIRCE1] Beams not factorizable: No simple power law: D B1 B 2 (x 1,x 2 ) 6= D B1 (x 1 ) D B2 (x 2 ) D B1 B 2 (x 1,x 2 ) 6= x 1 1 (1 x 1) 1 x 2 2 (1 x 2) 2 x e ± = x e ± CIRCE2 algorithm (WHIZARD 2.2.5, 02/15) Dalena/Esbjerg/Schulte [LCWS 2011] CLIC much more complicated (wakefields) Adapt 2D factorized variable width histogram to steep part of distribution Smooth correlated fluctuations with moderate Gaussian filter [suppresses artifacts from limited GuineaPig statistics Smooth continuum/boundary bins separately [avoid artificial beam energy spread]

17 Workflow GuineaPig/CIRCE2/WHIZARD

18 Workflow GuineaPig/CIRCE2/WHIZARD polarized spectra on demand

19 Iterations of Beam Spectrum (171,306 GuineaPig events in 10,000 bins)

20 Iterations of Beam Spectrum

21 NLO Development in WHIZARD Need for precision predictions that match (sub-) percent experimental accuracy Scary challenge for the theory community [ok, we have some time still ] Mostly electroweak corrections, but also QCD and pure QED Binoth Les Houches Interface (BLHA): Workflow 1. Process definition in SINDARIN (contract to One-Loop Program [OLP]) 2. OLP generates code (Born/NLO interference), WHIZARD reads contract 3. NLO matrix element loaded into WHIZARD

22 NLO Development in WHIZARD Need for precision predictions that match (sub-) percent experimental accuracy Scary challenge for the theory community [ok, we have some time still ] Mostly electroweak corrections, but also QCD and pure QED Binoth Les Houches Interface (BLHA): Workflow 1. Process definition in SINDARIN (contract to One-Loop Program [OLP]) 2. OLP generates code (Born/NLO interference), WHIZARD reads contract 3. NLO matrix element loaded into WHIZARD Working NLO interfaces to: GoSam [G. Cullen et al.] (first focus on QCD corrections) OpenLoops [J. Lindert et al.]

23 NLO Development in WHIZARD Need for precision predictions that match (sub-) percent experimental accuracy Scary challenge for the theory community [ok, we have some time still ] Mostly electroweak corrections, but also QCD and pure QED Binoth Les Houches Interface (BLHA): Workflow 1. Process definition in SINDARIN (contract to One-Loop Program [OLP]) 2. OLP generates code (Born/NLO interference), WHIZARD reads contract 3. NLO matrix element loaded into WHIZARD Working NLO interfaces to: GoSam [G. Cullen et al.] (first focus on QCD corrections) OpenLoops [J. Lindert et al.] WHIZARD v2.2.5 contains alpha version QCD corrections (massless and massive emitters) alpha_power = 2 alphas_power = 0 process eett = e1,e1 => t, tbar { nlo_calculation = full }

24 NLO Development in WHIZARD Need for precision predictions that match (sub-) percent experimental accuracy Scary challenge for the theory community [ok, we have some time still ] Mostly electroweak corrections, but also QCD and pure QED Binoth Les Houches Interface (BLHA): Workflow 1. Process definition in SINDARIN (contract to One-Loop Program [OLP]) 2. OLP generates code (Born/NLO interference), WHIZARD reads contract 3. NLO matrix element loaded into WHIZARD Working NLO interfaces to: GoSam [G. Cullen et al.] (first focus on QCD corrections) OpenLoops [J. Lindert et al.] WHIZARD v2.2.5 contains alpha version QCD corrections (massless and massive emitters) alpha_power = 2 alphas_power = 0 process eett = e1,e1 => t, tbar { nlo_calculation = full }

25 FKS Subtraction (Frixione/Kunszt/Signer) Subtraction formalism to make real and virtual contributions separately finite d NLO = Z n+1 d R d S {z } finite + Z Z d S + d V n+1 {z n } finite

26 FKS Subtraction (Frixione/Kunszt/Signer) Subtraction formalism to make real and virtual contributions separately finite d NLO = Z n+1 d R d S {z } finite + Z Z d S + d V n+1 {z n } finite Automated Subtraction algorithm: Find all singular pairs I = {(1, 5), (1, 6), (2, 5), (2, 6), (5, 6)} Partition phase space according to singular regions 1 = X 2I S ( ) Generate subtraction terms for singular regions

27 FKS Subtraction (Frixione/Kunszt/Signer) Subtraction formalism to make real and virtual contributions separately finite d NLO = Z n+1 d R d S {z } finite + Z Z d S + d V n+1 {z n } finite Automated Subtraction algorithm: Find all singular pairs I = {(1, 5), (1, 6), (2, 5), (2, 6), (5, 6)} Partition phase space according to singular regions 1 = X 2I S ( ) Generate subtraction terms for singular regions

28 Examples and Validation Simplest benchmark process: e + e! q q with NLO LO / LO = s / Plot for total cross section for fixed strong coupling constant List of validated QCD NLO processes e + e! q q e + e! q qg e + e! `+` q q e + e! `+ `q q e + e! t t Caveat: no fixed-order NLO event generation due to missing counter-event infrastructure Cross-checks with Madgraph5_aMC@NLO (except for ee WbWb) e + e! tw b e + e! W + W b b

29 Examples and Validation Simplest benchmark process: e + e! q q with NLO LO / LO = s / Plot for total cross section for fixed strong coupling constant List of validated QCD NLO processes e + e! q q e + e! q qg e + e! `+` q q e + e! `+ `q q e + e! t t e + e! tw b e + e! W + W b b Caveat: no fixed-order NLO event generation due to missing counter-event infrastructure Cross-checks with Madgraph5_aMC@NLO (except for ee WbWb) First working infrastructure for QCD NLO in pp First attempts on electroweak corrections, interfacing the RECOLA code [Denner et al.]

30 POWHEG Matching in WHIZARD Soft gluon emission before hard emission generate large logs Perturbative αs : M soft 2 1 k 2! log kmax T T kt min Matrix element + parton shower has to take this into account POWHEG method: hardest emission first [Nason et al.]

31 POWHEG Matching in WHIZARD Soft gluon emission before hard emission generate large logs Perturbative αs : M soft 2 1 k 2! log kmax T T kt min Matrix element + parton shower has to take this into account POWHEG method: hardest emission first [Nason et al.] Complete NLO events B( n )=B( n )+V ( n )+ Z d rad R( n+1 ) POWHEG generate events according to the formula: apple d = B( n ) NLO R (k min T )+ NLO R (k T ) R( n+1) B( n ) d rad Uses the modified Sudakov form factor: apple Z NLO R( n+1 ) R (k T )=exp d rad B( n ) (k T ( n+1 ) k T )

32 POWHEG Matching in WHIZARD Soft gluon emission before hard emission generate large logs Perturbative αs : M soft 2 1 k 2! log kmax T T kt min Matrix element + parton shower has to take this into account POWHEG method: hardest emission first [Nason et al.] Complete NLO events B( n )=B( n )+V ( n )+ Z d rad R( n+1 ) POWHEG generate events according to the formula: apple d = B( n ) NLO R (k min T )+ NLO R (k T ) R( n+1) B( n ) d rad Uses the modified Sudakov form factor: apple Z NLO R( n+1 ) R (k T )=exp d rad B( n ) (k T ( n+1 ) k T ) Hardest emission: kt max ; shower with imposing a veto: B<0 if virtual and real terms larger than Born: shouldn t happen in perturbative regions Reweighting such that B>0 for all events POWHEG: Positive Weight Hardest Emission Generator now implemented in WHIZARD

33 POWHEG Matching in e+e- to dijets

34 Top Threshold at lepton colliders ILC top threshold scan best-known method to measure top quark mass, ΔΜ ~ 100 MeV Heavy quark production at lepton colliders Threshold region (quantitatively)

35 Top Threshold in WHIZARD with F. Bach/A. Hoang/M. Stahlhofen Implement resummed threshold effects as effective tab vertex [form factor] in WHIZARD G v,a (0,p t,e+ i t, ) from TOPPIK code [Jezabek/Teubner], included in WHIZARD

36 Top Threshold in WHIZARD with F. Bach/A. Hoang/M. Stahlhofen Implement resummed threshold effects as effective tab vertex [form factor] in WHIZARD G v,a (0,p t,e+ i t, ) from TOPPIK code [Jezabek/Teubner], included in WHIZARD R,Z (s) = F v (s)r v (s) {z } s-wave: LL+NLL + F a (s)r a (s) {z } p-wave v 2 :NNLL BUT: differentially p-wave at NLL! Default parameters: M 1S = 172 GeV, t =1.54 GeV, s (M Z ) = Threshold/Continuum Matching: WIP

37 Top Threshold in WHIZARD with F. Bach/A. Hoang/M. Stahlhofen Implement resummed threshold effects as effective tab vertex [form factor] in WHIZARD G v,a (0,p t,e+ i t, ) from TOPPIK code [Jezabek/Teubner], included in WHIZARD R,Z (s) = F v (s)r v (s) {z } s-wave: LL+NLL + F a (s)r a (s) {z } p-wave v 2 :NNLL BUT: differentially p-wave at NLL! Default parameters: M 1S = 172 GeV, t =1.54 GeV, s (M Z ) = Threshold/Continuum Matching: WIP

38 Top Threshold in WHIZARD with F. Bach/A. Hoang/M. Stahlhofen Implement resummed threshold effects as effective tab vertex [form factor] in WHIZARD G v,a (0,p t,e+ i t, ) from TOPPIK code [Jezabek/Teubner], included in WHIZARD R,Z (s) = F v (s)r v (s) {z } s-wave: LL+NLL + F a (s)r a (s) {z } p-wave v 2 :NNLL BUT: differentially p-wave at NLL! Default parameters: M 1S = 172 GeV, t =1.54 GeV, s (M Z ) = Error estimate preliminary: DON T QUOTE!!! Threshold/Continuum Matching: WIP

39 Spin Correlation and Polarization in Cascades Cascade decay, factorize production and decay #evt/bin simulate (fullproc) #evt/bin simulate (casc) M inv (jl) M inv (jl) #evt/bin 800?diagonal_decay = true #evt/bin 1000?isotropic_decay = true M inv (jl) M inv (jl)

40 Spin Correlation and Polarization in Cascades Cascade decay, factorize production and decay #evt/bin simulate (fullproc) #evt/bin simulate (casc) M inv (jl) M inv (jl) #evt/bin 800?diagonal_decay = true #evt/bin 1000?isotropic_decay = true M inv (jl) M inv (jl) NEW: possibility to select specific helicity in decays! unstable W+ { decay_helicity = 0 }

41 Projects, Plans, Performance and all that O Mega Virtual Machine (OVM): ME via bytecode interpreter than compiled code Parton shower: LO merging (MLM ), NLO matching QED shower (FSR) QED shower (ISR); exclusive part of ISR spectrum pt spectrum of ISR radiation automated massless/massive QCD NLO corrections: FS / Initial state in preparation WHIZARD 3.0 QED/electroweak NLO automation: longer time scale complete NLL NRQCD top threshold/nlo continuum matching; extension to tth POWHEG matching implemented ; maybe also MC@NLO or Nagy-Soper matching Monte Carlo over helicities and colors Modified algorithm for multi-leg (tree) matrix elements: includes high-color flow amplitudes, QCD/EW coupling orders, general Lorentz structures Automatic generation of decays (and calculation of decay widths) New syntax for nested decay chains process = e1, E1 => (t => (Wp => E2, nu2), b), tbar

42 Conclusions & Outlook WHIZARD 2.2 excellent tool for Linear Collider Physics [ WHIZARD 2.2 excellent tool for LHC Physics] Great effort on the demands for mass production for LCs Beamspectra, LCIO, LC top threshold Main focus in physics: NLO automation WHIZARD 3.0 Performance: many developments to come Tell us what is missing, insufficient, annoying, desirable

43 Conclusions & Outlook WHIZARD 2.2 excellent tool for Linear Collider Physics [ WHIZARD 2.2 excellent tool for LHC Physics] Great effort on the demands for mass production for LCs Beamspectra, LCIO, LC top threshold Main focus in physics: NLO automation WHIZARD 3.0 Performance: many developments to come Tell us what is missing, insufficient, annoying, desirable even if it is in a LCWS summary talk Challenge accepted!

44 Conclusions & Outlook WHIZARD 2.2 excellent tool for Linear Collider Physics [ WHIZARD 2.2 excellent tool for LHC Physics] Great effort on the demands for mass production for LCs Beamspectra, LCIO, LC top threshold Main focus in physics: NLO automation WHIZARD 3.0 Performance: many developments to come Tell us what is missing, insufficient, annoying, desirable even if it is in a LCWS summary talk Challenge accepted!

45 (Personal) Memory to LCWS 2013: 金閣寺 J.R.Reuter Status Report on WHIZARD ALCW 2015, KEK,

46 J.R.Reuter Status Report on WHIZARD ALCW 2015, KEK,

47 ありがとうございます J.R.Reuter Status Report on WHIZARD ALCW 2015, KEK,